Establishing an Isolation Alert System for Mountain Community

The outreach road of mountain community has been interrupted by disasters such debris flow, flood and landslides, resulting in the interruption of the outreach road of the mountain community, forming a state like an island, which can be regarded as an isolation effect. In recent years, extreme events caused by extreme weather. The special geographical conditions in Taiwan, coupled with the increase in the frequency of natural disasters, have been heard by isolated island news. In 2015, Typhoon Soudelor hit Taiwan, and Wulai, New Taipei City caused severe disasters. Debris flow and landslides occurred, causing the interruption of Xinwu Road, the main liaison road in Wulai, and the isolation effect in Wulai. If we can integrate the historical data and research of isolation effect, and combine the theory of isolated prediction with instant rainfall and disaster prevention information, and finally visualize the information by alert system, it will help the general public's disaster prevention awareness and related disaster prevention unit decision-making reference.

Therefore, this research builds an isolation alert system. The three main information functions of this system include 1. disaster island geographic information function 2. isolated accident village identification function and 3. immediate isolated warning function. The d isolated geographic information display function is mainly to display the historical information about the isolation effect. The information of the village has been published, including the village's geography, social information and disaster history, and the risk map is presented by the vulnerability and resilience indicators. The village identification function of the isolated incident is realized by the Common Alerting Protocol of the road, and based on this, the identification in the immediate isolated village is carried out. The immediate disaster isolated warning function combines real-time rainfall information and integrates the Rainfall Triggering Index, Machine Learning's Supervised Learning algorithm, and the Common Alerting Protocol for the road. In the end, it was verified by the 2017 Typhoon Nepartak incident, and the results were all given the correct warning level for the isolated village.